Background Emergency medicine and paediatric hospital medicine physicians each provide a portion of the initial clinical care for the majority of hospitalised children in the USA. While these disciplines share goals to increase quality of care, there are scant data describing their collaboration. Our national, multihospital learning collaborative, which aimed to increase narrow-spectrum antibiotic prescribing for paediatric community-acquired pneumonia, provided an opportunity to examine factors influencing the success of quality improvement efforts across these two clinical departments.
Objective To identify barriers to and facilitators of interdepartmental quality improvement implementation, with a particular focus on increasing narrow-spectrum antibiotic use in the emergency department and inpatient settings for children hospitalised with pneumonia.
Methods We used a mixed-methods design, analysing interviews, written reports and quality measures. To describe hospital characteristics and quality measures, we calculated medians/IQRs for continuous variables, frequencies for categorical variables and Pearson correlation coefficients. We conducted in-depth, semistructured interviews by phone with collaborative site leaders; interviews were transcribed verbatim and, with progress reports, analysed using a general inductive approach.
Results 47 US-based hospitals were included in this analysis. Qualitative analysis of 35 interview transcripts and 142 written reports yielded eight inter-related domains that facilitated successful interdepartmental quality improvement: (1) hospital leadership and support, (2) quality improvement champions, (3) evidence supporting the intervention, (4) national health system influences, (5) collaborative culture, (6) departments’ structure and resources, (7) quality improvement implementation strategies and (8) interdepartmental relationships.
Conclusions The conceptual framework presented here may be used to identify hospitals’ strengths and potential barriers to successful implementation of quality improvement efforts across clinical departments.
- hospital medicine
- quality improvement
Statistics from Altmetric.com
Pneumonia is a common childhood infectious disease and a leading cause of paediatric hospitalisation, resulting in more than 100 000 hospitalisations and greater than US$500 million of healthcare costs annually in the USA.1 Global estimates suggest that children experience 120 million episodes of pneumonia each year, resulting in 1.3 million paediatric deaths.2 Recognising this substantial disease burden, broad efforts have been made to improve the quality of pneumonia care, including optimising antibiotic management and reducing unnecessary diagnostic testing and ineffective adjuvant therapies.3–7 However, the quality of care provided to children hospitalised with pneumonia varies substantially across hospitals.6 8–12 In the USA, provision of broad-spectrum antibiotics for childhood pneumonia, particularly ceftriaxone, is commonplace, with rates of ampicillin/penicillin prescribing less than 5% in several multisite studies prior to quality improvement (QI) efforts.11–13
Approximately 75% of children hospitalised in the USA for acute illnesses are cared for in emergency departments (EDs) prior to hospital admission, with initial disease management decisions correspondingly made by ED providers.14 Following initial evaluation and clinical management in EDs, care is usually handed off to paediatricians.15 16 Emergency medicine (EM) and paediatric training differ as do their attendant cultural mores, and there is scant literature describing the interface between these disciplines. In fact, differences in clinical training, practice culture and departmental affiliations may challenge effective collaboration.17 Few practice-based or evidence-based conceptual frameworks exist to inform interdepartmental QI collaboration in hospital settings, despite a widespread emphasis on QI within hospitals.
The Improving Community Acquired Pneumonia (ICAP) Collaborative, a multisite learning collaborative that aimed to increase narrow-spectrum antibiotic (NSA) prescribing for community-acquired pneumonia (CAP), provided an opportunity to examine QI implementation across clinical departments.18 Engaging healthcare providers across a national sample of hospitals, we aimed to identify barriers to and facilitators of interdepartmental QI implementation, with a particular focus on increasing NSA use in the ED and inpatient settings.
Most children in the USA are cared for at general hospitals (as opposed to free-standing children’s hospitals), where most ED physicians are not paediatric specialists. Approximately 80% of children hospitalised with CAP receive their initial care in the ED, with care then handed off to a paediatrician for the duration of the hospital stay.14
The ICAP Collaborative, conducted in 2015, was sponsored by the Value in Inpatient Paediatrics Network, part of the American Academy of Paediatrics (AAP) and was approved by the AAP Institutional Review Board (IRB). The primary goal of the collaborative was to increase appropriate NSA prescribing, with NSA defined as oral or intravenous amoxicillin, penicillin or ampicillin. Baseline NSA prescribing rates among participating hospitals were 14.7% (IQR 8.8%–27%) in the ED and 36.0% (IQR 21.2%–54.4%) in inpatient settings.18 Based on goals developed by a leadership team of CAP and QI experts, ICAP aimed to increase rates of NSA prescribing to >80% for eligible patients as detailed below.
Written informed consent was obtained from the team leader at each site, and local teams received IRB approvals as deemed necessary by each institution. Participating institutions were required to develop a multidisciplinary improvement team of at least three members. Inclusion of an EM physician on the team was encouraged but was not required. Project leadership facilitated educational webinars, a project listserv and individual site coaching by email and telephone. Each site was encouraged to implement a change package composed of (1) evidence-based pathways and order sets, (2) a toolkit for developing an antibiotic stewardship programme, (3) communication tools for promoting behaviour change and (4) slide sets for use in educational initiatives. Additional interventions to achieve project goals were implemented by individual sites based on locally perceived needs and facilitated by project leadership as well as coaching from national experts.
Children 3 months to 18 years of age hospitalised with a discharge diagnosis of pneumonia and who received antibiotics for CAP were considered for inclusion. Children with chronic, comorbid conditions predisposing them to recurrent respiratory illnesses, those requiring intensive care or a pleural drainage procedure, and those who underwent interhospital transfers were excluded. Each site had access to real-time run charts summarising their performance and was benchmarked by group aggregate performance quarterly. Full project methodology and results are published elsewhere.18
We employed a qualitative approach to identify barriers to and facilitators of collaboration for quality improvement between EM and paediatric hospital medicine (PHM) services, including content analysis of project reports collected during the ICAP Collaborative as well as semistructured interviews with ICAP site leaders following project completion. Site leaders were contacted by email to request their participation and informed consent for further interviews was obtained. First-round interviews focused broadly on sites’ local experiences with project implementation, while a second round of more focused interviews probed barriers to and facilitators of successful interdepartmental project implementation.
Our sampling framework for second-round interviews was informed by rates of NSA use achieved during the collaborative, with participants purposefully sampled using positive deviance sampling methods.19 We requested participation of sites achieving the highest rates of NSA prescribing in inpatient setting, including sites with both high and low rates of NSA antibiotic prescribing in the ED. This allowed for in-depth interviews of sites with concordant and discordant performance rankings between the inpatient and ED settings. All interviews were conducted by telephone with one member of the research team. Interviews were audiorecorded with permission, transcribed verbatim and verified for accuracy.
To describe hospital characteristics and rates of NSA prescribing, we calculated medians and IQRs for continuous variables, frequencies and percentages for categorical variables, and Pearson correlation coefficients to evaluate the relationship between rates of NSA prescribing in the ED and inpatient setting.
Using open-coding, an approach rooted in grounded theory, transcripts and written reports were reviewed by three members of the research team (CBA, ERO, JKL) to identify emergent concepts related to interdepartmental QI collaboration. Emergent concepts and associated definitions were summarised in a coding framework developed by the research team.20 21 Two members of the research team then independently applied codes to a subset of transcripts; areas of coding disagreement were resolved through in-depth discussions of the concepts, corresponding codes and definitions. Following assurance of coding agreement, transcripts were uploaded to Dedoose, a mixed-methods data management and analysis program, and one member of the research team coded the transcripts, with coding audits performed by the principal investigator.22 Analysis was performed during the interview period, and interviews were continued until the research team agreed that no new relevant concepts or insights were emerging from the data (data saturation).23–25 Following coding completion, all codes and associated transcript excerpts were reviewed by the research team to group similar concepts into themes, and similar themes into domains. During this axial coding process, domains were conceptualised within the context of the QI literature, with an aim to develop a framework to inform hospital-based interdepartmental QI implementation.
A total of 47 US-based hospitals participated in the ICAP Collaborative. Participating hospitals included community hospitals with paediatric beds, children’s hospitals nested within larger adult hospitals and free-standing children’s hospitals. Team composition varied substantially across sites; teams were led by paediatric hospitalists in partnership with nurses, pharmacists, EM physicians and other paediatric specialists (table 1). We analysed 142 written progress reports collected during three time points of the collaborative and conducted 35 semistructured interviews with 30 site leaders, including 27 first-round interviews and 8 second-round interviews.
Narrow-spectrum antibiotic prescribing
Figure 1 illustrates mean rates of NSA prescribing in the ED and inpatient setting achieved during the improvement phase of the collaborative. Median NSA prescribing was 43% (IQR 34%–55%) in the ED and 59% (IQR 43%–76%) in inpatient settings (p<0.01), with a moderately strong positive correlation between these rates in ED and inpatient services (r=0.49).
Barriers to and facilitators of interdepartmental QI implementation
We identified eight domains and associated themes regarding barriers to and facilitators of interdepartmental QI implementation. During our axial coding process, we identified parallels between our findings and Pettigrew and colleagues’ Receptive Contexts of Organizational Change framework.26 Therefore, building on the unique themes emerging in our research, we adapted Pettigrew’s framework of organisational change to illustrate our eight domains of interdepartmental QI in hospital settings and how these domains are inter-related (figure 2).
Hospital leadership and support
Within this domain, emergent themes included (1) hospital leadership and (2) administrative support for QI. Participants described how support from hospital leaders enabled more efficient implementation of changes to order sets and minimised bureaucratic barriers. In contrast, participants who reported less administrative support described substantial delays in ‘institutional roll out’. For example, one participant reported, “There is very poor oversight of, and little uniformity in, pathway development, approval and dissemination… We also struggled with getting our inpatient order set approved—the process has taken us nearly 5–6 months. This is because of the bureaucracy associated with creating/modifying order sets, despite the fact that our modification to the existing order set was minor” (P11). Participants reported how institutional leadership and support were closely related to departmental structure and resources, including systems to access and analyse administrative data and ‘protected time’ for project implementation. For example, one participant whose site achieved below-average rates of NSA prescribing stated, “If I’m not doing my clinical work I’m not getting paid… It wasn’t like (hospital administration) said ‘No, you can’t do it.’ … It wasn’t supportive but it wasn’t obstructionist” (P3). Institutional leadership and support was also described as influential to the organisational culture, with participants describing project participation as ‘in line with our mission and values’ (P16).
Quality improvement champions
Emergent themes within the domain of QI champions included (1) multidisciplinary team composition and (2) project advocates in each department. High performing sites reported high levels of engagement of residents, nurses, respiratory therapists, pharmacists and staff physicians from across departments. In the words of one participant, “We truly have a great team comprised of multiple representatives from all parts of the hospital—ED, pulmonary, infectious disease, and the residency programme. Each of the members of our team has been able to effectively engage his/her respective groups to help facilitate change” (P14). In contrast, other sites described significant challenges with multidisciplinary engagement and ‘holding each other accountable’ (P7) given the large number of disciplines involved with clinical management of pneumonia. Teams with high rates of NSA prescribing described the importance of having a champion within the ED to serve as a voice to their professional group, and that the absence of such a voice challenged effective information dissemination and changes to clinical practices. For example, “…I think a lot of ERs are like ours where they have a mix of different types of providers and it’s hard. As a hospitalist, you feel like you are preaching to the emergency room and I think having a champion in the ER is helpful” (P2). In addition, successful project implementation was described as being facilitated by infrastructure and personnel to support QI efforts, including hospital-wide QI staff and antimicrobial stewardship programmes. Accordingly, this domain was related to departments’ structure and resources, as well as the organisational culture.
Evidence supporting the intervention
Within this domain, emergent themes included (1) the role of evidence, (2) physician knowledge and preferences, and (3) competing demands. Participants described the importance of evidence to justify the value of the QI intervention and how physician preferences were, at times, at odds with the evidence base. One participant described hesitance by some physicians to follow evidence-based guidelines given concerns that they should not ‘supersede their clinical judgement’ (P25). Competing demands created by multiple concurrent QI initiatives were also reported as a barrier to project success, with one participant describing how implementation of another QI intervention encouraging rapid broad-spectrum antibiotic (BSA) administration challenged effective implementation of the current project: “One great barrier to the use of narrow spectrum antibiotics in the ED setting is the competing stop sepsis protocol. Any child with fever and tachycardia enters into the sepsis pathway, receiving Ceftriaxone as the antibiotic of choice. They are given a 24 hours dose, and so in many instances, it is the only antibiotic they receive before transfer to the inpatient floor” (P5).
National health system influences
Participants reported how national-level factors supported interdepartmental project implementation at the hospital level, describing the value of support from a national physicians’ organisation (AAP), national clinical guidelines to inform project goals and multihospital participation in the project. For example, one participant stated, “In my little community hospital, I don’t really have time for this kind of thing… But to me (being backed by a national organization) was one of the things that made it possible” (P6). This domain was described as being particularly important to smaller hospitals with limited past experience implementing QI efforts, with one participant stating, “I liked feeling that I was part of a bigger group trying to meet a common goal” (P25), and another stating, “It made me feel more in touch with the academic centers around” (P3).
Within this domain, emergent themes included (1) interdepartmental buy-in and (2) clinical practice culture, with participants describing different cultures and practice styles in the ED and inpatient setting. One site leader from a low-performing site stated, “You have different physicians, different providers, coming from different schools… so I wouldn’t say that everybody was on the same page… and I still do not say that everybody is on the same page right now” (P27). While some participants reported that both ED and PHM departments were ‘very open to changes and improvements’ (P8), others described how differences in the culture of the ED and inpatient unit challenged project implementation. The importance of local evidence was reported as integral to interdepartmental buy-in, with one participant stating, “I didn’t really feel that I could make the change in the ED to start with… If I could have done it routinely on the floor and shown that outcomes were good, then I think it would have been easier to sell to the ED. But to have gotten them to buy in initially probably would have been a challenge” (P7).
Perspectives shared by participants in this domain were closely related to the departmental structure domain. For example, siloed departmental structures often isolated ED and PHM teams from one another, with resulting differences in the clinical practice culture and pneumonia management priorities. Differing clinical priorities and approaches were particularly evident in general hospital EDs where care for paediatric patients comprised a small proportion of the overall patient volumes, and the majority of ED providers did not have pediatric-specific training.
Departmental structure and resources
Themes encompassed by this domain included (1) siloed departmental structures, (2) departmental size and (3) the potential of EMRs to facilitate interdepartmental change. With respect to siloed departments, one participant reported, “Politically it has been difficult, partly because the paediatric emergency department is part of the department of emergency medicine at our general hospital, and not the department of paediatrics” (P5). This challenge was echoed by many participants who described how having the ED and PHM groups in separate departments made it difficult to organise group meetings and to share information. Hospitalists working at larger hospitals more frequently discussed challenges with information dissemination to large departments with a shift-schedule-based workforce. For example, “The ER physicians group is massive. There are probably 30 some physicians alone in the emergency medicine group and then additionally there is mid levels, nurse practitioners and PAs. So trying to get that culture change was a bit more of a challenge” (P26). In contrast, participants from small hospitals reported opportunities to “utilise the strengths of being a smaller place, (including) personal relationships with people” (P3).
Participants at general hospitals (non-children’s hospitals) described specific challenges engaging EM physicians who did not have paediatric-specific training; this was one of the most frequently discussed challenges by participants. For example, one participant reported, “As a general hospital without board-trained paediatric emergency physicians, we need to help them be more comfortable and capable of managing paediatric illnesses… that’s probably by biggest goal…” (P20).
Quality improvement implementation strategies
Within this domain, participants described (1) successes and challenges associated with their QI implementation strategies, (2) the value of multidisciplinary education and (3) time constraints for QI implementation. Varied approaches to multidisciplinary education were described by participants, including personalised report cards, ‘email blasts’, formal educational sessions to residents and ‘face-to-face education to outlying sister facilities and community medical providers’ (P14). Participants reported inconsistent success with flyers and printed order sets, but consistently described the value of small group education. For example, one participant reported, “It seemed like education in small groups with key providers worked better than any sort of widespread system changes, so these small sort of grassroots efforts rather than wide process efforts ended up being the key…” (P22). Implementation of EMR notifications and electronic order sets, when supported by hospital infrastructure, was frequently endorsed as effective by participants, illustrating the inter-relatedness of this domain with that of departmental resources.
Themes within this domain included (1) direct conversation between departments and (2) the role of ‘boundary spanners’.26 Participants described how participation in the learning collaborative forged new and previously undeveloped relationships between departments, with one stating, “We recognise that broad spectrum coverage started in the ED, almost always. So the head of the ED group and the head of the hospitalist group brought both groups around a table and we talked, for the first time ever at our institution” (P14). Participants also emphasised the role of residents, fellows and physicians as ‘boundary spanners’ whose work across hospital settings enabled change across these settings, as well as the value of effective relationships between staff physicians across departments. For example, “The team includes a highly motivated paediatric resident… This was paramount since residents work in multiple areas of the hospital and are often the ones putting in orders for patients. They also serve a vital role in the ED where many of the attendings are not peds ED trained” (P14). Participants also described the value of specific physicians whose training, moonlighting experience, cross-appointments or personal relationships enabled successful project implementation across departments.
In this mixed-methods study of a national collaborative to improve paediatric CAP management, we identified barriers to and facilitators of collaboration for QI across two hospital departments. Key facilitators of change included institutional support, QI champions and interdepartmental relationships, particularly facilitated by ‘boundary spanners’ such as trainees and physicians whose work spanned multiple clinical settings. In contrast, key barriers to successful collaboration included siloed clinical departments, challenges with effective information dissemination through large clinical departments, and competing QI priorities. Our conceptual framework, informed by the organisational change literature, fills a gap in the published literature about factors that support or hinder effective QI efforts across clinical departments in hospitals.
Among both children and adults admitted to hospital with general medical conditions such as CAP, the majority experience hospital-based care in multiple departments and from multiple healthcare providers.27 28 Despite this, formal efforts to improve healthcare processes and outcomes across clinical departments are relatively rare, perhaps because sponsoring organisations are often discipline or disease specific. As a result, few conceptual frameworks exist to guide cross-departmental QI efforts at the hospital level. The framework resulting from our research was adapted from the work of Pettigrew and colleagues who, in 1992, published an analysis of organisational change within England’s National Health Services, derived from qualitative analysis of eight District Health Authorities.26 In this work, which examined change within a national system, they coined the phrase ‘receptive contexts for change’ and summarised eight drivers of organisational change.
The themes emerging in our work are unique to interdepartmental QI efforts within a hospital, but the overarching domains mirror the structure of Pettigrew’s work in several areas, with some notable differences. Pettigrew’s eight receptive contexts for change are summarised in table 2, adjacent to our own findings. While Pettigrew highlighted the importance of co-operative interorganisational networks for national health system change, our analysis identified the importance of interdepartmental relationships within health systems. Within both of these domains, ‘boundary spanners’ were identified as playing an important role, facilitating change through existing relationships as well as knowledge of multiple systems or departments. The other notable difference between Pettigrew’s framework and our own was within the domain described by Pettigrew as ‘environmental pressure’. While their work identified financial pressures and managerial inertia as key factors influencing organisational change, participants in our study described health system influences beyond their own institutions, and specifically the positive benefits of participation in a multi-institutional learning collaborative sponsored by a national organisation.
Consistent with our findings, the importance of interdisciplinary relationships to improve healthcare quality has been called out by a number of professional societies. For example, effective interdisciplinary communication is a key component of efforts to improve hospital-to-home transitions for hospitalised patients and is called out in a policy statement endorsed by seven professional organisations.29–31 In addition, the importance of interdisciplinary relationships has been the focus of a number of studies of nurse–physician interactions.32–34 There is less information about improving physician–physician relationships and communication, particularly when different disciplines are involved. In this study, paediatricians felt that belonging to a different discipline or department was a significant barrier to interdisciplinary QI efforts with EM physicians. In settings where faculty and/or residents crossed disciplines, this was felt to be a facilitator.
One unique barrier emerging from our work was the importance of considering ‘competing demands’ when multiple QI initiatives are implemented concurrently in hospitals. Participants reported that BSAs were frequently initiated in the ED; this is evidenced by the lower rate of NSA prescribing observed in EDs in our study, as well as in past studies of pneumonia management.9 Several potential reason for this emerged in our research, including our finding that NSAs recommended for CAP management are at odds with the BSAs recommended for sepsis management.35 Given widespread implementation of QI initiatives within US hospitals, poor performance on quality measures may result from multiple, uncoordinated projects, differing priorities and differing practice cultures across departments. Co-sponsorship of QI initiatives by multiple clinical departments at the time of project planning may address some of these barriers, and is a lesson learnt from this project.
Our results should be interpreted in light of this study’s strengths and limitations. Strengths of our approach include the inclusion of a large national sample of hospitals and our ability to conduct two rounds of interviews to probe more deeply about interdepartmental implementation. In addition, the availability of quantitative data allowed us to purposefully sample participants based on performance within the collaborative. The inclusion of participants from both children’s hospitals and general community hospitals is also a strength, as community hospitals are typically under-represented in published QI work. Study limitations include the fact that only site leaders, all of whom were paediatric hospitalists, were interviewed, excluding the perspectives of key stakeholders in other disciplines and departments. In addition, we acknowledge that potential for response bias and coding bias may have influenced our results, although we made efforts to minimise these by conducting interviews following completion of the collaborative, by double-coding a subset of transcripts and by ensuring participants that their responses would be de-identified.
In conclusion, this work presents a framework to recognise and address barriers to and facilitators of interdepartmental QI in structurally diverse hospitals. We build on existing organisational change literature, highlighting unique contextual factors that should be considered during project planning and implementation of QI efforts across clinical departments in hospitals. Recognising that the vast majority of patients who are admitted to hospital receive care in both the ED and inpatient settings, consideration of lessons learnt in this work may improve the quality and efficiency of healthcare delivery and the success of future interdepartmental QI efforts.
Contributors All of the authors are responsible for this research; all have participated in the concept and design, analysis and interpretation of data, drafting or revising the manuscript, and have approved the manuscript as submitted.
Funding Dr Leyenaar was supported by grant number K08HS024133 from the Agency for Healthcare Research and Quality. The content is solely the responsibility of the authors and does not necessarily represent the official views of the Agency for Healthcare Research and Quality.
Competing interests None declared.
Patient consent Not required.
Ethics approval American Academy of Pediatrics.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.